Abstract There is a disproportionate burden of stroke among African Americans (AA) and individuals in the Southeastern United States, although the environmental and/or genetic underpinnings remain incompletely defined. Because metabolites integrate both environmental and genetic factors, they serve as proximal markers of human disease. In prior work supported by a Career Development Award, we established a rapid, multifunctional metabolomics platform that can be applied to large epidemiological cohorts. We now seek to focus our study on the metabolite predictors of cerebrovascular disease, an area that has received relatively little attention. To address these gaps in knowledge, we propose conducting metabolomics in stroke in the multiethnic REasons for Geographic and Racial Differences in Stroke ( REGARDS) study, with further replication in the Jackson Heart Study (JHS) and Framingham Heart Study (FHS). We will test the overarching hypothesis that metabolomics in well-phenotyped populations will illuminate stroke-associated metabolite pathways, and unravel racial differences among them. In Aim 1, we will identify metabolite risk factors for incident stroke, stratified by subtype, in the case-control cohort of the REGARDS study. We will profile metabolites from the baseline examination in ~1200 incident stroke cases and ~1200 controls matched for age, sex and race. In Aim 2, we will identify novel metabolite markers of known clinical risk traits for stroke, including atrial fibrillation, hypertension, diet and exercise. In Aim 3, we will validate novel metabolite predictors of incident stroke and associated risk traits in the JHS and FHS cohorts. We will also test whether genetic variants that determine metabolite levels are in turn associated with clinical traits, through genetic risk score analysis. With extensive study data, the REGARDS study is uniquely positioned to catalyze the Aims of this proposal. Leveraging a novel small molecule profiling platform, our preliminary studies demonstrate the feasibility and significance of the Aims. Our multidisciplinary collaboration includes investigators at the MGH, UAB and University of Cincinnati, who bring collective expertise in metabolite profiling, biomarkers, genetic and population epidemiology, bioinformatics, metabolic traits, and health disparities. Finally, all data will be made publicly available, producing a unique scientific resource for the stroke research community.